This fix adds PIDs 9801 and 9802 to the pybcdc.inf file.

When in CDC only mode, it presents itself as a Communcations
device rather than as a composite device. Presenting as a
composite device with only the CDC interface seems to confuse
windows.

To test and make sure that the correct pybcdc.inf was being used,
I used USBDeview from http://www.nirsoft.net/utils/usb_devices_view.html
to uninstall any old pyboard drivers (Use Control-F and search
for pyboard). I found running USBDeview as administrator worked best.

Installing the driver in CDC+MSC mode first is recommended (since the
pybcdc.inf file in on the internal flash drive). Then when you switch
modes everything seems to work properly.

I used https://github.com/dhylands/upy-examples/blob/master/boot_switch.py
to easily switch the pyboard between the various USB modes for testing.

Not remove so far, may be needed later.

This patch eliminates the need for a nested parse node for assignments
and keyword arguments.  It saves a little bit of RAM when parsing.

In Python 3.4 the value is evaluated before the key.  In Python 3.5 it's
key then value.

Also adds tests specifically for testing constant folding.

Thanks to Peter Hinch.  Addresses issue #1488.

OpenJDK seemed to return NULL in case of exception, but Dalvik returns
arbitrary value, so skip such "optimizations".

Thanks to Peter Hinch.  See issue #1490.

* Move the esp.status() to network module.
* Describe the wifi.isconnected() method.
* Describe esp.mac(), esp.wifi_mode(), esp.phy_mode(), esp.sleep_type(),
  esp.deepsleep(), and esp.flash_id() functions.

From https://github.com/pfalcon/axtls , branch micropython.

All breaking changes going from 3.4 to 3.5 are contained in
basics/python34.py.

With this patch parse nodes are allocated sequentially in chunks.  This
reduces fragmentation of the heap and prevents waste at the end of
individually allocated parse nodes.

Saves roughly 20% of RAM during parse stage.

Python semantics are that rhs of shift must be non-negative, so there's
no need to handle negative values in the underlying mpz implementation.

This patch adds more fine grained error message control for errors when
parsing integers (now has terse, normal and detailed).  When detailed is
enabled, the error now escapes bytes when printing them so they can be
more easily seen.

To avoid local ref table overflow.